By James Kwak
(Warning: long post ahead.)
I was minding my own business, reading Past Due by Peter Goodman (I got it from Simon, who I think got it for free), and there on page 43 I ran into Eric Bochner. I thought that name sounded familiar, and then I remembered what it was. Eric Bochner was a vice president of something or other (and then the vice president of something else or other) at Ariba, where I worked from April 2000 until September 2001 (I was also a consultant there from December 1999). Chapter 2 of Goodman’s book is about the Internet bubble, Ariba is his case study, and Bochner is his source.
As far as I know, no one has made Ariba the poster child for the Internet bubble before–people usually go with WebVan, or Pets.com, or something similarly vaporous. Ariba is a more complicated story, but you can make a case that we deserve to be on the poster. At our peak we were bigger than all those pet food companies combined, with a market capitalization over $40 billion (on quarterly revenues of about $100 million at that time). More to the point, if Pets.com is comedy, Ariba is tragedy (well, not really, but you know what I mean): Ariba was a real company with a real product that got swept up in its own hype, with unfortunate consequences (but not fatal ones–Ariba today earns over $300 million in annual revenues and a small profit).
Goodman wants to make the case that there was no there there, and Bochner says what he needs him to say.
It was never about creating a business, he said. ‘The whole idea was attract investment, sell your story to the Street, and then get yours out’ (p. 47).
But there was a there there, at least at the beginning. Ariba’s founders–mainly experienced veterans of other successful technology companies, not twenty-somethings with no clue–had a great idea (automating procurement by large companies), and for the first few years almost everything went well. (Ariba was founded in 1996 by seven people; Bochner joined in October 1999, two months before I showed up as a consultant.)
The product really worked, it was better than anything out there, the deals were huge, and most of the customers were happy. Goodman says,
there were obvious limitations to how much money Ariba could make selling its software. It was aiming its product at the big Fortune 500 companies. What happened when Ariba ran out of customers? (pp. 44-45).
But you can build an enormous company selling only to Fortune 500 companies and their global equivalents (that’s where the money is, after all), and even during the frothy days of the boom we sold almost exclusively to big companies (because that’s where the money is).
In Goodman’s account, Ariba was basically a fraud. In the book, Bochner was
the lead man in creating [the impression that Ariba was constructing a global marketplace]. . . . From the beginning, Bochner saw this as a rather impossible task,
but he went through the motions anyway. Why? In Goodman’s words,
the stock was the only thing that mattered. A valuable stock gave Ariba currency it could use to buy other companies (p. 46).
As I recall, Ariba started out very much as a real company, but was actually blindsided by the Internet boom. (Update: Note that I only joined Ariba in 2000, so my knowledge of the early days is based on talking to many people who were there, not actually being there.) In 1999, as industry and financial analysts were talking up the business-to-business e-commerce boom, they were looking for the leaders.
The problem was that since business-to-business e-commerce barely existed, there were no leaders. Ariba was the closest thing they could find, at that was already a stretch (at the time, Ariba’s product did a good job of automating the process of creating a purchase order, but most purchase orders were automatically faxed to suppliers). Ariba had recently launched the Ariba Network, which was a hub where suppliers could publish electronic catalogs and buyers could route their purchase orders to suppliers, and the analysts hungrily anointed it the next big thing.
There was nothing particularly wrong with the Ariba Network; it did what we told the customers it did, more or less. The problem was the hopes that were projected onto it by analysts and, increasingly, by Ariba itself. Ariba’s stock price kept climbing and climbing, to the point where Ariba had to at least try to deliver on the market’s expectations; otherwise, its stock would have collapsed and it would have been acquired. We got into a vicious cycle where the more the stock climbed, the more dependent we were on the story that was making it climb, and the more ambitious and aggressive we had to be.
We were by no means innocent victims. There was too much watching the stock price, and a lot of believing our own marketing. The jargon that Goodman makes fun of was terrible, and I was responsible for some of it. More important, no one put a gun to the CEO’s head and said,
you must try to build a global network of electronic marketplaces and earn transaction fees by providing services to all the participants in that marketplace.
The management team could have said,
This is silly. We’re going to ignore the hype and focus on what we do well, which is automating corporate procurement of indirect goods
(not direct manufacturing inputs). That’s what we should have done, even though it would have meant a collapse in the stock price and being acquired. But the problem is we were swept up in the hype like everyone else. It wasn’t just the analysts. The customers all wanted to be building global electronic marketplace, and in that group I include just about every large corporation in the world; they were all talking to us and to our competitors (Commerce One, Oracle, i2). They all wanted an e-commerce strategy (because without one their stock prices were being pounded) and we were their e-commerce strategy. The customers came to us and said,
We want this, everyone says you’re the best, therefore you’re our best hope. And here’s a big pile of money.
And we did our best to build what they wanted. The problem wasn’t that the software was impossible to build; we could have built it, although it would have taken a lot longer than anyone wanted to wait. The problem was, as Bochner says, the degree of industrial reorganization required was just not feasible in the short timeframe. It was a business model problem: there was no business model for, say, a single giant marketplace where small business bought everything they needed from a universe of suppliers, all hosted by Bank of America (BAC). Bank of America thought there was gold in that hill, and we sold them a semi full of shovels.
Back in early 2000, I think most of us thought it was possible. Otherwise, it’s hard to explain the energy, the motivation, and the immense amount of work most of us did. Goodman says on page 55,
The tech start-ups and their Wall Street handlers implicitly ridiculed the work ethic as an antiquated notion while speaking to a different lobe of the American brain–the part ruled by the frontier dream of finding a mountain of gold and never working again.
I beg to differ. Maybe the dream was to never work again, but most people realized that to get there you had to work very, very hard, and often at multiple startups. (There are certainly examples of people who used connections to parachute into startups just before they went public, but those people are generally hated within the startup world.) Goodman is right that the discount brokers like Ameritrade did create the illusion that you could become rich by sitting in your pajamas and trading stocks, but I don’t think the image of Silicon Valley ever included the idea of knocking off work at noon and hitting the golf course (except maybe for venture capitalists). If Ariba really had been a fraud, then the only people working hard would have been in marketing, PR, and investor relations.
Over time, as with all bubbles, we began to doubt. By late 2000, I think most of us realized that reaching the expectations the market had for us was a long shot. And as I said, the fundamental problem wasn’t the software, although the software wasn’t as good as it could have been. It was that the customers were failing at building the marketplace communities as fast as they had envisioned, and as the bubble imploded suddenly they were no longer interested anymore, and suddenly the market expected us to collapse. As it turned out, the customers didn’t really want what they had been desperately scrambling for less than a year before. And we had over two thousand people lined up to build shovels that no one wanted.
There are obviously parallels to the financial craze of 2004-2007, and we did more or less play the part of Chuck Prince (music, dance, etc.). All of the paper wealth went to the heads of people at the boom companies, who started thinking they were smarter and better-looking than they actually were, just like overpaid bankers are convinced they are actually smarter than their college friends who got Ph.D.s in physics and are now struggling junior professors. (Unlike the Wall Street bubble, however, most of that wealth stayed on paper, because of the rule that stock options vest over four years, and the tendency for most people not to cash out at the first opportunity.)
But the big difference, as many have noted, is that ours was an equity bubble, not (for the most part) a debt bubble. Technology startups were funded by venture capital firms, which generally have no debt, and by stock offerings; they generally don’t have the assets and steady cash flows to borrow money. (The telecom infrastructure companies were an exception.) There was little margin borrowing to buy stocks, as compared to the 1920s, so when the NASDAQ crashed, the damage did not spill over into other asset classes, and the financial system didn’t even wobble.
In Goodman’s story, the tech bubble had a scarring effect on the American psyche–two of them, actually. On the one hand, it fueled the myth that you could get rich through investments (buying tech stocks), which shifted household behavior away from working and saving and toward living off of anticipated capital gains. On the other hand, he says the collapse of the tech bubble created a "paralyzing state of cynicism” (p. 81) that hurt the economy for years afterward. Well, which one is it?
I think the simplest explanation is that people invested in tech stocks because they wanted to get rich, and they got burned; the real estate bubble is considerably more complicated, because in many cases people had to overpay for housing, because people have to buy houses (or rent, but in many areas it’s hard to find good rentals–that’s a whole other story), whereas no one had to buy stock in WebVan.
For those ex-Aribians wondering what happened to Eric Bochner who don’t want to buy the book, he’s apparently now running a chocolate factory in Iowa City, and he implies that he didn’t take much money off the table when he could (“The money I had then, I wish I had now” [p. 49]). I do know a few people who took everything off the table that they could, but only a very few. Most people lost their paper wealth along with everyone else’s paper wealth (and unlike Dick Fuld, they hadn’t sold hundreds of millions of dollars of stock on the way up).
